> ## Documentation Index
> Fetch the complete documentation index at: https://docs.dataharbor.co/llms.txt
> Use this file to discover all available pages before exploring further.

# Field Targeting

> Target nested fields and arrays consistently across Data Control and Data Transform.

# Field Targeting

Field paths are shared across [Data Control](./data-control) and [Data Transform](./data-transform). Once DataHarbor matches an object definition, it resolves each field path against the normalized payload for that object.

<Info>
  Object matching happens first. Field targeting happens second. See [Virtual APIs](../core/virtual-apis) and [Data Pipeline](../core/data-pipeline) for the higher-level request flow.
</Info>

## Quick reference

| Path                     | Targets                                                 |
| ------------------------ | ------------------------------------------------------- |
| `email`                  | A top-level scalar field                                |
| `user.email`             | A nested field inside an object                         |
| `"[contacts].email"`     | The `email` field on every item in the `contacts` array |
| `"[orders].[items].sku"` | A nested field inside nested arrays                     |

## Path syntax rules

* Use dot notation for nested objects.
* Use bracket segments like `"[contacts]"` for arrays.
* Slash notation like `address/city` is not supported.
* If the matched payload itself is a root array, the full controls pipeline runs against each object element automatically.

## Nested fields

Use dot notation when the field lives inside one or more nested objects.

```yaml theme={null}
controls:
  - type: redact
    fields:
      - user.ssn
      - billing.card.last4
      - metadata.created_by
```

## Arrays

Use bracket segments when the field exists inside every item of an array.

```yaml theme={null}
controls:
  - type: tokenize
    fields:
      - "[contacts].email"
      - "[contacts].phone"
      - "[orders].[items].sku"
```

If the response payload itself is a root array, you do not need a leading array segment. DataHarbor applies the control list to each object element automatically.

## `required: true`

Set `required: true` when a missing field should fail the request instead of being skipped.

```yaml theme={null}
controls:
  - type: hash
    fields: [customer_id]
    required: true
```

Required checks are full-path strict:

* If an intermediate parent object is missing, the request fails.
* If the final field is missing, the request fails.
* For `coalesce`, all referenced fields being missing or empty also fails when `required: true` is set.

## Scalar versus structured targets

Filter-style data controls such as `redact`, `tokenize`, `anonymize`, `mask`, `hash`, and `allow` target scalar values only.

That means:

* strings, numbers, and booleans are valid targets
* objects and arrays are not valid targets for filter controls

If you need to reshape objects instead of replacing scalar values, use [Data Transform](./data-transform).

## Transform path rules

Transforms add a few extra path constraints:

* For `combine` and `coalesce`, every source field in `fields` plus the destination `into` path must share the same parent path.
* `delete` does not use `into`; it removes the paths named in `fields` directly.
* If `into` already exists, `combine` and `coalesce` fail instead of overwriting the field.

## Next steps

<CardGroup cols={2}>
  <Card title="Data Control" icon="shield" href="./data-control">
    Apply privacy controls to the fields you target
  </Card>

  <Card title="Data Transform" icon="shuffle" href="./data-transform">
    Create, combine, and remove fields in the same pipeline
  </Card>

  <Card title="Control Block Reference" icon="code" href="../api-reference/control-block-reference">
    Review the detailed rules for each control type
  </Card>
</CardGroup>